status-report-drafter-scott-margetts

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.75). The skill’s runtime workflow ingests the user-provided “unstructured input (pasted emails, call notes, Teams messages, verbal summaries)” and extracts substantive updates from it; if that pasted content includes outsider-authored free text (e.g., counterparty/team messages), that text is read and incorporated into the LLM context for report drafting.